sign in sign up
<<
Home , MRPL23, MRPL24, MRPL37, MRPS7, Translation (biology)

Neurobiology of Disease 141 (2020) 104880

Contents lists available at ScienceDirect

Neurobiology of Disease

journal homepage: www.elsevier.com/locate/ynbdi

A homozygous MRPL24 causes a complex movement disorder and affects the mitoribosome assembly T

Michela Di Nottiaa,1, Maria Marcheseb,1, Daniela Verrignia, Christian Daniel Muttic, Alessandra Torracoa, Romina Olivad, Erika Fernandez-Vizarrac, Federica Moranib, Giulia Trania, Teresa Rizzaa, Daniele Ghezzie,f, Anna Ardissoneg,h, Claudia Nestib, Gessica Vascoi, Massimo Zevianic, Michal Minczukc, Enrico Bertinia, Filippo Maria Santorellib,2, ⁎ Rosalba Carrozzoa, ,2 a Unit of Muscular and Neurodegenerative Disorders, Laboratory of Molecular , Bambino Gesù Children's Hospital, IRCCS, Rome, Italy b Molecular Medicine & Neurogenetics, IRCCS Fondazione Stella Maris, Pisa, Italy c MRC Mitochondrial Unit, University of Cambridge, Cambridge, UK d Department of Sciences and Technologies, University Parthenope of Naples, Naples, Italy e Unit of Medical and Neurogenetics, Fondazione IRCCS Istituto Neurologico Carlo Besta, Milan, Italy f Department of Pathophysiology and Transplantation, University of Milan, Milan, Italy g Child Neurology Unit, Fondazione IRCCS Istituto Neurologico Carlo Besta, Milan, Italy h Department of Molecular and Translational Medicine DIMET, University of Milan-Bicocca, Milan, Italy i Department of Neurosciences, IRCCS Bambino Gesù Children Hospital, Rome, Italy

ARTICLE INFO ABSTRACT

Keywords: Mitochondrial ribosomal large 24 (MRPL24) is 1 of the 82 protein components of mitochondrial ribo- Mitochondrial disorders somes, playing an essential role in the mitochondrial process. Movement disorder We report here on a baby girl with cerebellar atrophy, choreoathetosis of limbs and face, intellectual dis- MRPL24 ability and a combined defect of complexes I and IV in muscle biopsy, caused by a homozygous missense mu- Mitoribosomes tation identified in MRPL24. The variant predicts a Leu91Pro substitution at an evolutionarily conserved site. Mitochondrial protein synthesis Using human mutant cells and the zebrafish model, we demonstrated the pathological role of the identified Zebrafish fi fi Molecular modeling variant. In fact, in broblasts we observed a signi cant reduction of MRPL24 protein and of mitochondrial Protein interactions respiratory chain complex I and IV subunits, as well a markedly reduced synthesis of the mtDNA-encoded . In zebrafish we demonstrated that the orthologue is expressed in metabolically active tissues, and that induced locomotion impairment, structural defects and low ATP production. The motor phenotype was complemented by human WT but not mutant cRNA. Moreover, sucrose density gradient frac- tionation showed perturbed assembly of large subunit mitoribosomal , suggesting that the mutation leads to a conformational change in MRPL24, which is expected to cause an aberrant interaction of the protein with other components of the 39S mitoribosomal subunit.

1. Introduction The human MRP gene family comprises 30 encoding for the small mitochondrial ribosomal subunit and 52 genes for the large subunit, all Mammalian Mitochondrial Ribosomal Proteins (MRPs) are encoded of which are nuclear encoded (Gopisetty and Thangarajan, 2016). by nuclear genes and are required for the translation of the 13 mtDNA- Building a mitoribosome is a difficult logistical task comparable to the encoded proteins within the . Mitochondrial complexity of assembling the entire respiratory chain, since the bio- (mitoribosomes) consist of a small 28S subunit and a large 39S subunit. genesis of mitoribosomes depends on the coordinated synthesis of

⁎ Corresponding author at: Unit of Muscular and Neurodegenerative Disorders, Laboratory of Molecular Medicine, Bambino Gesù Children's Hospital, IRCCS, Viale di San Paolo 15, 00146 Rome, Italy. E-mail address: [email protected] (R. Carrozzo). 1 These authors contributed equally to this work. 2 These authors contributed equally as senior authors. https://doi.org/10.1016/j.nbd.2020.104880 Received 12 July 2019; Received in revised form 4 March 2020 Available online 25 April 2020 0969-9961/ © 2020 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

MRPs, which must be translated on cytoplasmic ribosomes and im- glutamine supplemented with sodium pyruvate (0.11 g/L) 10% fetal ported into mitochondria (Bogenhagen et al., 2018). Mitoribosomes are bovine serum, and 50 μg/mL uridine. Complex V activity (in the di- essential for the synthesis of the oxidative machinery. rection of ATP synthesis) was measured in fibroblast mitochondria, It has indeed been proposed that the absence or deficiency of MRPs may using reported spectrophotometric methods (Rizza et al., 2009). cause primary oxidative phosphorylation disorders (O'Brien et al., Oxygen consumption rate (OCR) was measured in adherent fibroblasts 2000). The consequences of in MRP genes are expected to with an XFe24 Extracellular Flux Analyzer (Seahorse Bioscience,Agi- reproduce the clinical heterogeneity of oxidative phosphorylation dis- lent, Santa Clara, CA), as described in Tolomeo et al., 2019. Western orders, ranging from lethality to less severe phenotypes associated with blotting using monoclonal antibodies complex I – NDUFA9 and marginally impaired energy metabolism, as mutations that do not fully NDUFS1; complex II – SDHA; complex III – UQCRC2; complex IV – inactivate the protein function may have intermediate effects (O'Brien COXIV; complex V – ATP5A1, and porin (VDAC) was performed using et al., 2005). In addition, mutations in MRP genes could cause tissue- precast 4–12% denaturating gels (Bolt™ 4–12% Bis-Tris Plus Gels, specific disorders (Spirina et al., 2000). The crucial role of mitoribo- Thermo Fisher Scientific, Waltham, MA USA) as reported (Torraco somes in OXPHOS biogenesis underlines the involvement of MRPs in a et al., 2018). All monoclonal antibodies were from MitoSciences (Eu- number of conditions associated with mitochondrial dysfunction (Bieri gene, OR, USA). The polyclonal ab-MRPL24 was from Abcam (Cam- et al., 2018). The importance of the proper composition and structure of bridge, UK). mitoribosomes for human health has led to their recent structural Analysis of mitochondrial protein synthesis was performed as pre- characterization, which provided detailed insight in the protein-rich viously described (Chomyn, 1996; Fernández-Silva et al., 2007). Im- mammalian mitoribosome (Amunts et al., 2015; Greber et al., 2015), mortalized fibroblasts at 70% confluence were labeled for 1 h with and a structure-based kinetic model for its assembly (Bogenhagen et al., [35S]-L- in the presence of 100 μg/mL emetine, an inhibitor 2018). Finally, very recently, dysfunction of these proteins has been of cytosolic protein synthesis. Samples were kept at −80 °C until use. demonstrated to play a role not only in the primary mitochondrial re- Twenty μg of total cellular protein were loaded on a Novex™ 18% Tris- spiratory chain activity deficiencies, but also in other conditions, such Glycine precast SDS polyacrylamide gel (Invitrogen). The gel was then as , impaired development, neurodegeneration, cardiovascular fixed and dried, and the mitochondrial translation products were vi- failure, obesity and inflammatory disorders (Gopisetty and sualized using a Storage Phosphor Screen and a Typhoon 9410 Variable Thangarajan, 2016). Mode Imager (GE Healthcare). Mitochondrial large 24 (MRPL24) is one of the For all experiments, age-matched controls were used. protein components of the large (39S) subunit of mitochondrial ribo- somes (Kenmochi et al., 2001). Herein, we report on the first case with 2.4. Model building and in silico analysis mutation in MRPL24 and provide a comprehensive investigation of its clinical, biochemical and structural consequences. We demonstrated For the analysis of the MRPL24 structure in the context of human the pathological role of the identified variant using human mutant cells. mitoribosome, the structure of the entire mitoribosome solved by Moreover, we evaluated the consequence of a dysfunctional Mrlp24 in electron microscopy at a resolution of 3.5 Å was used (PDB ID: 3j9m, the developing zebrafish embryos, in order to determine the impact on Amunts et al., 2015). As the coordinates of the MRPL24 159–168 early stages of development. fragment are missing from this structure, they were comparatively modeled with Modeller9v17 (Sali and Blundell, 1993), using as tem- 2. Material and methods plate the structure of the 89% sequence identical homologous protein from Sus scrofa (PDB ID: 4v19, Greber et al., 2014a, 2014b). To simu- 2.1. Standard protocol approvals, registrations, and patient consent late the structural effect of the Leu91Pro mutation, Leu91 was mutated to Pro and its phi (ϕ) dihedral angle was set at −63° with PyMol The study was approved by the Ethical Committees of the Bambino (DeLano, 2002): 63° is indeed the ϕ value typically assumed by prolines Gesù Children Hospital, Rome, Italy, in agreement with the Declaration in proteins (with variations within ± 15°, MacArthur and Thornton, of Helsinki. Informed consent was signed by the parents of the patient. 1991). The analysis of the interface areas and of the inter-residue contacts between MRPL24 and other proteins/ on the mitoribo- 2.2. Mutational analysis some was carried out with CoCoMaps (Vangone et al., 2011).

Total genomic DNA underwent targeting resequencing (BGI- 2.5. Analysis of the mitochondrial profile on sucrose density Shenzhen, Shenzhen, China) using a customized probe library (Agilent gradients SureSelectXT Custom Kit) designed to capture the coding exons and 20 of flanking of 1381 genes coding for mitochondrial Sedimentation and analysis of the was proteins (“Mitoexome”)(Calvo et al., 2012). Deep sequencing used an performed essentially as described in (Rorbach et al., 2016). Briefly, Illumina Hiseq technology, with a 255× effective mean depth. The control or Patient MRPL24 immortalized fibroblast cells were lysed and Burrows-Wheeler Aligner (BWA, http://bio-bwa.sourceforge.net/) loaded equally on separate linear sucrose gradients (10–30%) in 50 mM software was applied for analysis, classification, and reporting of Tris–HCl (pH 7.2), 20 mM Mg(OAc)2, 100 mM NaCl, and centrifuged genomic variants. After excluding previously annotated single nucleo- for 2 h 15 min at 39,000 rpm. Fractions of 100 μL each were collected, tide changes occurring with high frequency in populations (> 1%), we and 30 μL of each fraction was analyzed by western blotting using prioritized variants predicted to have a functional impact (i.e., non- mtLSU antibodies. The fractions within the blots were quantified using synonymous variants and changes affecting splice sites). Validation of ImageJ and normalized to the overall Control/Patient intensity ratio. the MRPL24 variant and segregation along the family was performed by Sanger Sequencing, using BigDye chemistry 3.1 and run on an ABI 2.6. Functional studies in zebrafish 3130XL automatic sequencer (Applied Biosystems, Technologies). Adult male and female wild-type (WT) zebrafish (AB strains) and 2.3. Human samples, biochemical and protein studies transgenic line expressing GFP under a strong cardiac , Tg (cmlc2:gfp) (Huang et al., 2003) were maintained according to standard Quadriceps muscle biopsy was obtained for diagnostic spectro- procedures (Westerfield, 2000) on a 14 h light:10 h dark cycle. Zeb- photometric procedures (Bugiani et al., 2004). Human fibroblasts, ob- rafish embryos and larvae procedures complied with the guidelines of tained from skin biopsy, were grown in high glucose DMEM with our institutional care committee, and experiments were

2 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 1. Brain MRI and genetic features of the MRPL24 patient. a. Brain MRI performed at age 3 years: (I) Axial T2 weighted image; (II) Axial FLAIR weighted image; (III, IV) Coronal T2 weighted images. High intense T2 and FLAIR abnormalities corresponding to the putamen bilaterally are evident in a, b and c. Cerebellar atrophy is shown in d. b. Electropherograms of the genomic region of the patient (Pt) and the parents showing the c.272T > C variant in exon 2 of MRPL24. c. Protein sequence alignment (ClustalW) highlights the mutation (p.Leu91Pro; red) in the human MRPL24. The Leu91 is conserved in the invertebrate and vertebrate or- thologs. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) performed in accordance with, and under the supervision of, the In- OR) targeting either translation start site (tblockMO) or at stitutional Animal Care and Use Committee (IACUC) of the University exon 4–5 splice site (spliceMO). The MO and primers sequences are of Pisa and the Italian National Research Council Institute of Clinical listed in Supplementary Table 1. Concentrations of MOs were carefully Physiology (CNR-IFC). Every effort was made to minimize both animal titrated to avoid nonspecific binding effects and a scrambled control suffering and the number of needed to collect reliable scientific MO was used at similar concentrations to assess specificity to . data. After titration, we used in all experiments 3 ng of spliceMO against Total RNA was extracted using the Quick RNA miniprep (Zymo mrpl24. The effect of the MO was evaluated through RT-PCR. Survival Research, Irvine, USA) from embryos at 48 h post fertilization (hpf) and was calculated comparing death rates between microinjected and non- reversely transcribed using the Transcriptor First Strand cDNA injected embryos. Rescue experiments were performed through co-in- Synthesis Kit (Roche, Hamburg, Germany). Primers for mRNA se- jection of 50 pg of control, mrpl24 and either human WT or mutant quences were designed using the zebrafish sequence of mrpl24 (mutLeu91Pro) cRNAs with spliceMO at the same concentration used for (ENSDART00000010862.6), and the human transcript the knockdown experiments. Each experiment was repeated at least (ENST00000361531.6 1). PCR products were cloned into pCS2+ vector three times if not otherwise stated. and verified for accuracy using capillary sequencing. Birefringence assay in 48 hpf embryos was used to test muscle Whole mount in situ hybridization (WISH) was performed as de- structure and compaction (Smith et al., 2013). Heart development was scribed (Thisse and Thisse, 2008) at the 48, 72, and 96-hpf time points. analyzed using the Tg(clmc2:GFP) embryos anesthetized with 0.04 mg Antisense riboprobe synthesis was performed using the Digoxigenin of tricaine (E1052, Sigma) and set down on glass slides embedded in (DIG) RNA Labeling Kit (Roche). Quantitative reverse transcription 1.2% low melting agarose (Agarose low gelling, Sigma). The fluores- followed by the polymerase chain reaction in real time (qRT-PCR) ex- cence imaging was carried out using the Nikon Eclips E600 microscope, pression analyses was done as described (Marchese et al., 2016). and acquired with CoolSnap-CF camera using NIS elements software For the generation of zebrafish mrpl24 knockdown, we designed version 2. For whole mount staining, dechorionated embryos were fixed antisense (MOs) (GeneTools, Philomath, in 4% PFA overnight and Oil-Red-O (ORO) staining was performed as

3 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 2. Decreased levels of MRPL24 and OXPHOS subunits in MRPL24 patient muscle. a, b. Muscle homogenate from controls (Ctrls, n = 3) and MRPL24 patient (Pt) was separated on a 4–12% SDS-PAGE and proteins were visualized by western blot analysis using specific antibodies against MRPL24. The same samples were tested for the expression of subunits of the CI (NDUFA9), CIII (UQCRC2), CIV (COXIV), CV (F1β). The levels of the different proteins were normalized to VDAC (corre- sponding histograms). Data are presented as a mean ± SD of at least three independent experiments; *p < .05; **p < .001; ***p < .0001. described (Fraher et al., 2016). Images were acquired on a Leica M80 3. Results microscope with a Nikon Digital Sight DS-Fi1 camera and the NIS Elements software package (Nikon, Nikon Corp., Europe). 3.1. Clinical features and brain NMR findings Touch-evoked escape response was measured at 48 hpf on a semi- quantitative arbitrary scale from 0 to 2, where 0 is no movement, 1 is The patient was the single daughter of healthy unrelated parents. flicker of movement but no swimming, and 2 is normal swimming Motor milestones were reported normal until the age of 8 months when (Marchese et al., 2016). she almost acquired sitting position, but within a few weeks, she lost Western blotting in zebrafish used a polyclonal anti-MRPL24 the ability of sitting with support, and vocalisation. She was admitted to (1:500), an anti-Total Oxphos (1:1000, Mitoprofile), an anti-complex I our Hospital for the first time at 3 years of age, and clinically she was (1:1000, Mitosciences), and anti-β-tubulin (1:1000, Signaling) as not able to sit with support and was lying in supine position, showing control. spasticity and dystonic posturing of distal upper and lower limbs to- Oxygen consumption rates in 48-hpf control embryos and mor- gether with dyskinetic facial grimacing. At that time, the MRI showed phants (n = 10 per group) were measured using an extracellular flux T2 and FLAIR hyperintense bilateral alterations at the level of the pu- analyzer (XFe24, Seahorse Biosciences) as described (Gibert et al., tamen and cerebellar atrophy (Fig. 1A). The child was evaluated again 2013). Single embryos were placed in individual wells of a 24-well islet at the age of 8 years and was able to sit with support; but she had plate and a fine screen mesh was used to maintain the larvae in place. intellectual disability with lack of interaction with the environment and All the sequences of the primers and morpholino used are available aspects of a pervasive behavior. Neurological examination at this age in the Supplementary table 1. showed choreoathetosis of limbs and face. Increased lactate was de- tected in blood 2.2 mM (n.v.: 0.9–1.8 mM). Respiratory chain com- plexes (RCCs) analysis in muscle biopsy revealed multiple defects 2.7. Statistical analysis [RCCIV (−62%), RCCI (−40%) and RCCIII (−42%)]. A second biopsy was performed 3 years later and confirmed the RCCIV defect (−58%), All data were analyzed applying either parametric or non-para- while the other RCC activities were in the control range. Molecular metric analyses. Homogeneity of variance was assessed using the genetic analysis in the muscle tissue excluded deletions and depletion of Levene test. Post hoc comparisons were performed by Mann-Whitney the mitochondrial DNA. The girl was re-assessed at the age of 14 years test with Bonferroni's correction, or Unpaired t-test following non- and, although she was relatively stable, she could not stand and was parametric analysis of variance. ANOVA with Tukey's Multiple wheelchair bound; choreoathetosis persisted with spastic-dystonic tet- Comparison Test and Fisher's exact and Chi-square test were used in raparesis. Although the patient was not able to speak, her interaction specific zebrafish experiments. with the environment had improved, particularly after she had been put For functional studies on human samples, data are presented as in treatment with Idebenone at the dose of 180 mg/day. The electro- mean ± SD. The Student's t-test was used for the analysis of statistical cardiogram disclosed a RS of 91 bpm, ventricular pre-excitation with significance. associated repolarization anomalies, QTc within limits; the child never had symptoms of a paroxysmal tachycardia in the clinical history. Echocardiogram was normal, though a Holter ECG confirmed findings

4 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 3. Decreased levels of MRPL24 and OXPHOS subunits, as well mitochondrial protein synthesis in MRPL24 patient fibroblasts. a, c. Fibroblasts homogenate from controls (Ctrls, n = 2) and MRPL24 patient (Pt) was separated on a 4–12% SDS-PAGE and proteins were visualized by western blot analysis using specific antibodies against MRPL24. The same samples were tested for the expression of subunits of the CI (NDUFA9; NDUFS1), CII (SDHA), CIV (COXIV). The levels of the different proteins were normalized to VDAC (corresponding histograms). Data are presented as a mean ± SD of at least three independent experiments; *p < .05; **p < .001; ***p < .0001. b. Mitochondrial translation study in two controls and the mutant immortalized fibroblasts cell lines. The experiment was repeated twice in two different batches of cells showing the same result. suggestive of a Wolff-Parkinson-White syndrome without any other 3.3. Functional studies on human samples arrhythmia. She is currently on therapy with Riboflavin 200 mg/day, Carnitine 1 g/day, and Idebenone 180 mg/day with the recommenda- The level of MRPL24 protein was dramatically reduced in muscle tion of careful cardiac monitoring. homogenate (Fig. 2a) associated with a decreased level of subunits of complexes I, IV and V, when normalized with the mitochondrial marker VDAC (Fig. 2b). The fibroblasts showed a significant reduction of the 3.2. MitoExome and Sanger sequencing amount of MRPL24 protein as well (Fig. 3a). Specific 35S labelling of the mtDNA-encoded peptides revealed markedly reduced synthesis of all of Bioinformatic analysis carried out on a MitoExome targeted panel, them in the mutant immortalized fibroblasts (Fig. 3b), and overlapping on the hypothesis of a recessive trait, identified biallelic mutations in a results were obtained measuring the amount of MRC subunits by wes- single gene, MRPL24 (NC_000001.11). Sanger sequencing confirmed tern blot (Fig. 3c). In addition, biochemical analysis on fibroblast mi- the presence of the c.272T > C (p.Leu91Pro) variant in a homozygous tochondria documented a defect of complex V activity using either state in the patient and in heterozygosity in her healthy parents succinate, malate or pyruvate+malate as substrate (Fig. S1). Using (Fig. 1b). The mutation was located in exon 2 and was not reported in microscale oxygraphy to evaluate the impact of the mutation on mi- public databases (dbSNP (http://www.ncbi.nlm.nih. gov/sites/), ExAC tochondrial respiration we observed significantly reduced basal and (http://exac.broadinstitute.org/), EVS (http://evs.gs.washington.edu/ maximal respiratory capacities, together with a low mitochondrial ATP EVS/), and gnomADv2.1 (http://gnomad.broadinstitute.org), and in in- output (Fig. S2). house databases. In silico analysis with Polyphen-2 (PPH2, http:// genetics.bwh.harvard.edu/pph2/), SIFT (http://sift.jcvi.org), and Mu- tation Taster (http://www.mutationtaster.org) predicted a harmful ef- 3.4. In silico characterization of MRPL24 mutation fect of the p.Leu91Pro variant which affects a residue conserved along all the invertebrate and vertebrate phyla (Fig. 1c). Human MRPL24 is a 216 protein, highly interconnected with other components of the human mitoribosomal large subunit. It spans a length of over 15 nm on the surface of the large 39S

5 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 4. Three-dimensional representation of the entire human mitoribosome (PDB ID: 3j9m). MRPL24 and all the protein/RNA components it interacts with are shown in a surface representation and labeled. For 16S RNA, only the surface of nucleotides 1800–1810 is shown, for the sake of clarity. The rest of the ribosome is shown in a cartoon representation, with proteins colored pink if in the small and lightblue if in the large subunit and RNAs colored hotpink in the small and deepblue in the large subunit. Inset: MRPL24 is shown in the same orientation as in the mitoribosome. The hypothetical orientation of the beta hairpin motif involving residues Thr93-Glu112, in the Leu91Pro mutant is also shown, colored in hotpink. It was obtained by setting the value of the ϕ dihedral angle of residue 91 to −63° that is the typical accepted value for prolines. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) mitoribosome subunit, where it interacts with five proteins, uL29m ribosomal large subunit in the patient fibroblasts. Upon western blot (MRPL47), mL41 (MRPL41), uL23m (MRPL23), mL45 (MRPL45) and analysis of whole cell homogenates, we observed decreased levels mi- mL37 (MRPL37), and the 16S RNA, sharing with them a total interface toribosomal proteins MRPL41 and MRPL45 (both of which are posi- area as large as ≈ 5100 Å2 (Fig. 4, S3). tioned adjacent to MRPL24) in patient cells when compared to control Leu91 in MRPL24 is located at the basis of a beta-loop-beta (or beta cells (Fig. 5a-b). Similarly, upon sucrose density gradient fractionation, hairpin) structural motif encompassing residues Thr93-Glu112. This we observed decreased levels of MRPL24, MRPL41 and MRPL45 in the motif provides extensive contacts with the protein mL45 (MRPL45) and fractions corresponding to the large subunit of the mitoribosome the 16S RNA (specifically with its residues 1800–1810) (Fig. S3). The (Fig. 5c-e, fractions 6–8). This suggests the large subunit is structurally value assumed by the Leu91 ϕ angle in MRPL24 (−147°) is clearly compromised by the Leu91Pro mutation in MRPL24, which may explain incompatible with the presence of a Pro, which has the ϕ angle re- the observed phenotype and poor mitochondrial translation activity. stricted around −63° ( ± 15°) (MacArthur and Thornton, 1991). This implies that the Leu91Pro mutation will cause a conformational change 3.6. Functional studies in zebrafish on the protein main chain that will likely affect the orientation of the following beta hairpin motif (Fig. 4) and downstream. The aberrant MRPL24 in human and Danio rerio presents about 70% of identity folding of mutated MRPL24 may lead to reduced steady-state levels, for both protein and mRNA (Fig. 1c) with a high degree of synteny. By consistent with the above data (Fig. 2a and 3a). < it is expected to qRT-PCR studies, we observed that mrpl24 had the highest levels of result in a less efficient interaction of mutant MRPL24 with other expression at 72 hpf (Fig. 6a) and it was expressed until 120 hpf, albeit components of the mitoribosome, especially mL45 (MRPL45) and 16S to a lesser extent. WISH showed mrpl24 mRNA signals ubiquitously rRNA. A defective interaction between its molecular components can expressed at 48hpf (Fig. 6b), whereas expression was restricted to so- cause, in turn, a disturbance in the 39S mitoribosomal subunit as- mites, brain, heart, and liver at 96 hpf (Fig. 6c-e), suggesting a higher sembly. expression in metabolically active tissues. Using a specific spliceMO, we observed minimal death rate and no 3.5. Mitoribosome integrity analysis morphological changes until 48 hpf when MO-injected embryos showed smaller size, less pigmentation, a slight curvature of the body, and Next we set out to analyze the integrity of the mitochondrial occasionally, heart edema (Fig. 7a) associated with a reduction of

6 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 5. Effect of MRPL24 mutation on the mitochondrial ribosome. a. Western blot analysis of fibroblast homogenate from control (Ctrl, n = 3) and MRPL24 patient (Pt, n = 3) samples. b. Quantification of band intensities in (a) using ImageJ normalized to Beta . Data are presented as mean ± SEM; *p < .05. c-e. Sucrose density gradient fractionation and western blot analysis of mitoribosomal large subunit proteins in Ctrl and Pt samples. Quantification of band intensities using ImageJ normalized against the fraction of the total signal relative to control cells. mRNA expression and protein level (Fig. S4a-b). Over 80% of 48 hpf and Ban, 2016). As a rule, the small subunit is implicated in the de- morphants showed altered touch evoked response test, and unlike the coding of mRNAs, while the large subunit carries out the peptidyl- wild-type could not run-off the visual field in a longer registration of activity for formation (Bieri et al., 2018). In recent their locomotion (Fig. 7b, video S1,2). Contrary to WT larvae, 120 hpf years, most of and mammalian mitoribosomal proteins have been morphants were slowly moving or paralyzed (video S3,4). The altera- identified and further confirmed as components of the mitochondrial tions of the locomotion at 48hpf were rescued after the co-injection of ribosome by high-resolution cryo-electron microscopy structures (De the WT zebrafish mrpl24 cRNA (Fig. 7c). A similar effect was obtained Silva et al., 2015; Greber and Ban, 2016). Moreover, it has been re- by complementing the touch evoked response phenotype of the mor- ported that these proteins not only have a structural role but also bio- phants with WT human but not with the mutLeu91Pro cRNA, suggesting logical functions in translation (Brodersen and Nissen, 2005), even if that the Leu91Pro allele is likely deleterious and acts with a loss of the exact function of many of them is not yet clear. A hallmark of the function mechanism (Fig. 8). mitochondrial ribosome is that, despite the proteobacterial origin, Like human mutant cells, 48 hpf mrpl24 morphants showed reduced during evolution, certain ribosomal RNA segments have been lost and basal respiration rates and impaired ATP production compared to WT replaced by mitochondria-specific ribosomal proteins (Smits et al., siblings by microscale-oxygraphy (Fig. 9a,b) and had low expression of 2007; Desmond et al., 2011) and this structural change was accom- MRC complexes I and V when tested by Western blotting (Fig. S5). To panied by the acquisition of specialized functions for the synthesis of a assess how knocking-down mrpl24 would impair organ developments, small set of highly hydrophobic mitochondrial inner membrane pro- we injected 3 ng spliceMO in a transgenic zebrafish line expressing GFP teins (Ott and Herrmann, 2010). under the promoter of myl7, at the level of cardiomyocytes and per- Of the 82 proteins components of the human mitoribosome, up to formed morphological examination of heart structure and birefringence now only 10 have been linked to mitochondrial diseases: MRPS2 assay. While we did not observe obvious alterations in birefringence (Gardeitchik et al., 2018), MRPS7 (Menezes et al., 2015), MRPS16 assay (data not shown), 72 hpf morphant embryos presented a slight (Miller et al., 2004), MRPS22 (Saada et al., 2007; Smits et al., 2011; but significant increase of cardiac edema with an elongated and string- Baertling et al., 2015), MRPS23 (Kohda et al., 2016), MRPS25 like heart tube, a feature not appearing in WT siblings (Fig. S6). The (Bugiardini et al., 2019), MRPS34 (Lake et al., 2018) of the mt-SSU and cardiac phenotype was complemented by co-injection of MO with the MRPL3 (Galmiche et al., 2011), MRPL12 (Serre et al., 2013) and mutLeu91Pro cRNA (Fig. 9c). Hepatic size was also increased in 5 dpf MRPL44 (Carroll et al., 2013; Distelmaier et al., 2015) of the mt-LSU. morphants with a significant level of lipid storage, a finding suggesting Defects in these proteins cause heterogeneous and multi-systemic the presence of liver steatosis (Fig. 9d,e). clinical phenotypes, ranging from features common to almost all the described cases, which include severe lactic acidosis, combined OX- PHOS deficiency, hypotonia and developmental delay, to specific clin- 4. Discussion ical presentations such as cardiomyopathy (Galmiche et al., 2011; Smits et al., 2011; Carroll et al., 2013; Distelmaier et al., 2015), neurodeve- Here we report the first case of a patient harboring a mutation in the lopmental disabilities (Miller et al., 2004; Distelmaier et al., 2015; Lake MRPL24 gene, which codes for a structural protein of the mitochondrial et al., 2018), corpus callosum agenesis (Saada et al., 2007), Leigh large ribosomal subunit. Mitoribosomes comprise a 28S small subunit syndrome (Lake et al., 2018), hypoglycemia (Kohda et al., 2016; (SSU) made of 12S rRNA and about 30 proteins, and a 39S large subunit Gardeitchik et al., 2018), or growth delay and dysmorphic features (LSU) composed of 16S rRNA and approximately 52 proteins (Greber

7 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 6. a. qRT-PCR analysis showing the expression of mrpl24 until 120 hpf, with highest levels of expression at 72 hpf. b. In situ-hybridization reavealing that mrpl24 mRNA is ubiquitously expressed at 48hpf, whereas, expression was restricted to somites (*, sm), brain (*), heart (*, h), intestine (int), and liver (*, liv) at 96 hpf (c-e), compared to the sense probe (f).

(Saada et al., 2007; Smits et al., 2011; Lake et al., 2018; Gardeitchik mitochondria-specific protein mL45 (MRPL45) and extensions of et al., 2018). Here we describe a patient presenting hyperintense bi- uL23m (MRPL23), uL24m (MRPL24), and uL29m (MRPL29) (Brown lateral alterations at the level of the putamen, cerebellar atrophy and et al., 2014; Greber and Ban, 2016). mL45 (MRPL45) likely anchors the choreoathetosis of limbs and face with a spastic-dystonic tetraparesis. mitoribosome to the inner mitochondrial membrane and exposes the Moreover, the electrocardiogram disclosed a RS of 91 bpm and ven- translated nascent polypeptide to solvent (Greber et al., 2014a, 2014b; tricular pre-excitation with associated repolarization anomalies. RCC Brown et al., 2014). The mutation we characterized in MRPL24 is im- analysis in muscle biopsy revealed a defect of complex IV activity, as- mediately upstream to a conserved beta hairpin structural motif that is sociated with decreased level of complexes I, IV and V subunits. The in close contact with mL45 (MRPL45) (Brown et al., 2014). A proline in amount of MRPL24 protein was reduced both in muscle and fibroblasts, this positionis incompatible with the structure of the wild-type protein, causing markedly reduced levels of the assembled mitochondrial ribo- thus necessarily leading to an aberrant folding and, consequently, to an some large subunit and reduced translation of all the mtDNA-encoded aberrant interaction of MRPL24 with other components of the large subunits. These results provided evidence of the deleterious effect of the mitoribosomal subunit. Overall, our data suggests that the Leu91Pro p.Leu91Pro MRPL24 variant in the mitochondrial translation process. mutation destabilizes MRPL24, leading to reduced incorporation into To support this, a structural analysis in the context of the human mi- the mitoribosome and preventing its efficient interaction with neigh- toribosome has been performed. MRPL24 is highly interconnected with bouring proteins. As a consequence, the mitochondrial large subunit is other proteins and with the 16S RNA within the LSU subunit con- also destabilized. A similar effect was observed for the analogous pa- tributing, together with mL41 (MRPL41), uL23m (MRPL23), uL29m thogenic mutation Leu214Pro in MRPS2, hypothesized to cause a (MRPL47) and bL34m (MRPL34), to the formation of the peptide exit conformational change leading to a decrease in the steady-state levels tunnel (Amunts et al., 2015; Bogenhagen et al., 2018). The exit tunnel of mutant MRPS2, and preventing the assembly of the small mitor- of the mammalian LSU subunit differs from that of the bacterial 50S ibosomal subunit (Gardeitchik et al., 2018). subunit due to the loss of two rRNAs and the addition of the In addition, we used an in vivo vertebrate system (zebrafish) to

8 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 7. a. 48 hpf WT and MO-injected embryos, showing the smaller size, decreased pigmentation, slight curvature of the body, and, occasionally, heart edema in the MO. b. Touch evoked response test reveal that over 80% of 48hpf morphants (n = 361) showed altered escape response and , unlike the wild-type could (n = 488). c. Morphant phenotype rescue after the co-injection of the zebrafish mrpl24 mRNA (MO n = 165; MO + WT mrpl24 cRNA n = 144). Each bar represents the mean + s.e.m.; ***p < .0001; n.s., not significant according to ANOVA with Tukey's Multiple Comparison Test (b) or Unpaired t-test (c).

considerable reduction of the basal respiration and ATP production, highlighting the presence of mitochondrial dysfunction associated to faulty mrpl24. The heterogeneity of MRPs patients is well represented in our model as zebrafish embryos exhibited reduced development, he- patomegaly and lipid accumulation (Howarth et al., 2013; Vacaru et al., 2014). Moreover, we found that MRPL24 has a potential role in cardiac development producing heart impairment. In fact, it has been reported that some MRPs defects gave rise to cardiomyopathy (Galmiche et al., 2011; Smits et al., 2011; Carroll et al., 2013; Distelmaier et al., 2015)or to a large patent ductus arteriosus (Miller et al., 2004). The relatively milder cardiac phenotype observed in our patient could be due to the different mutation (missense in patient, knock-down in zebrafish). However, overexpression of the p.Leu91Pro mutation rescues the heart phenotype of the morphants, but not the locomotion, suggesting that the mutation impacts mainly on motor behavior and less on cardiac function. Conversely, liver involvement, a feature not seen in our pro- band, could be related to impaired synthesis of ATP in zebrafish directly influencing fatty acid oxidation, steatosis, cell death, and fibrogenesis (Lee and Sokol, 2007). Whilst we cannot speculate on the appearance of liver symptoms in our patient with disease progression, other mutations affecting MRPs have been associated with liver dysfunction (Menezes et al., 2015; Kohda et al., 2016; Lake et al., 2018). In conclusion, our study demonstrates for the first time that MRPL24 is required for normal mitochondrial function in zebrafish and Fig. 8. a. Analysis of the touch evoked response test in uninjected (n = 183), humans, and that mutations destabilizing this protein underlie the morphants (n = 261) and MO co-injected with human WT cRNA (n = 71), and clinical manifestations of mitochondrial dysfunction. with human Leu91Pro cRNA (n = 262), showing that the mutated mRNA harboring the Leu91Pro mutation did not rescue the phenotype differently than the WT cRNA. ***p < .0001, n.s., not significant according to Chi-square test. Role of authors

M. Di N., C. D. M., G. T., T. R., E. F-V., C. N.: performed functional explore how dysfunctional Mrpl24 interferes with mitochondrial bioe- analysis on muscle and fibroblasts. nergetics and with organs development, as well as to explore how the M. M., F. M.: created the morphant zebrafish and performed related identified missense mutation in this gene underlies the pathological functional analysis. effects. In the mrpl24 knock-down zebrafish embryos, we observed D. V., A. T.: performed NGS, Sanger experiments and the related motor impairment and a metabolic damage associated with a bioinformatics analyses.

9 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Fig. 9. a-b. Micro-oxygraphy analysis of 48 hpf mrpl24 morphants (n = 29) and control uninjected (n = 29) embryos, showing the reduction of the basal respiration rates and impaired ATP production in morphants. c. The cardiac phenotype was complemented by co-injection of the human mutant Leu91Pro cRNA (n = 150). d-e. Morphological examination of heart 72hpf morphants (n = 180) presented a slight but significant increase of cardiac edema with an elongated and string-like straight heart tube not appearing in WT siblings (n = 115). Moreover, Oil-Red-O staining and in situ hybridization with fabp10a, showed significant level of lipid accumulation indicative of steatosis and increased liver size, respectively in morphants. ***p ≤ .0001,**p < .001, n.s., not significant according to Mann–Whitney test (a-b) and to Fisher exact Contingency test (c). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

R. O.: provided model building and analysis on the structure of the M. M., R. O., D. G. M. Z., Mi. M, E. B., F. M. S., R. C.: Funding mutated protein. acquisition. D. G.: performed biochemical study on muscle samples and revised the manuscript. Declaration of Competing Interest A. A., G. V., E. B.: Acquisition and analysis of clinical data. M. Di N., M.M., R. O., R.C.: write the original draft. None of the authors has any conflicts to disclose. M. Z., Mi. M, E. B., F.M. S., R. C.: conceptualization of this study, drafting and editing of the manuscript, analysis and interpretation of Acknowledgements data. fi All authors: reviewed and edited the nal version of the manuscript. EB and RC are supported by the Italian Ministry of Health Ricerca Corrente; EB is supported also by the Italian Ministry of Health Ricerca CRediT author statement Finalizzata (RF-2016-02361241). FMS was partially supported by a grant from Regione Toscana (RR/NUTRA-FISH). DG is supported by the The corresponding author ensures that the descriptions of the au- Telethon Foundation (GGP15041). RO thanks MIUR-FFABR (Fondo per thors' role is accurate and have been shared with all authors. il Finanziamento Attività Base di Ricerca) for funding. M.M. was par- M. Di N., M. M., D. V., C. D. M., A. T., R. O., E. F-V., F. M., G. T., T. tially supported by the Starting Grant from Italian Minister of Health R., C. N.,: Methodology; Formal analysis. Ricerca Finalizzata. M.Z. was supported by the Core Grant from the M. Di N., M. M., D. V., C. M., A. T., R. O., E. F-V., F. M., G. T., T. R., MRC (Grant MC_UU_00015/5); ERC Advanced GrantFP7-322424 and A. A., C. N., G. V.: Investigation. NRJ-Institut de France Grant. Mi.M. and C.D.M. were supported by the M. Di N., M.M., R. O., R.C.: writing the original draft. Core Grant from the MRC (Grant MC_UU_00015/4). We also thank Dr. D. G., M. Z., Mi. M, E. B., F. M. S., R. C.: Writing - Review & Editing. Letizia Pitto for technical and scientific assistance at the Italian R. C.: Supervision. National Research Council (NRC)-zebrafish core facility.

10 M. Di Nottia, et al. Neurobiology of Disease 141 (2020) 104880

Appendix A. Supplementary data Huang, C., et al., 2003. Germ-line transmission of a myocardium-specific GFP transgene reveals critical regulatory elements in the cardiac myosin light chain 2 promoter of zebrafish. Dev. Dyn. 228, 30–40. Supplementary data to this article can be found online at https:// Kenmochi, N., et al., 2001. The human mitochondrial ribosomal protein genes: mapping doi.org/10.1016/j.nbd.2020.104880. of 54 genes to the and implications for human disorders. 77, 65–70. Kohda, M., et al., 2016. A comprehensive genomic analysis reveals the genetic landscape References of mitochondrial respiratory chain complex deficiencies. PLoS Genet. 12, e1005679. Lake, N.J., et al., 2018. Biallelic mutations in MRPS34 lead to instability of the small Amunts, A., et al., 2015. Ribosome. The structure of the human mitochondrial ribosome. mitoribosomal subunit and Leigh Syndrome. Am. J. Hum. Genet. 102, 713. Science 348, 95–98. Lee, W.S., Sokol, R.J., 2007. Liver disease in mitochondrial disorders. Semin. Liver Dis. Baertling, F., et al., 2015. MRPS22 mutation causes fatal neonatal lactic acidosis with 27, 259–273. brain and heart abnormalities. Neurogenetics 16, 237–240. MacArthur, M.W., Thornton, J.M., 1991. Influence of proline residues on protein con- Bieri, P., et al., 2018. High-resolution structures of mitochondrial ribosomes and their formation. J. Mol. Biol. 218, 397–412. functional implications. Curr. Opin. Struct. Biol. 49, 44–53. Marchese, M., et al., 2016. Dolichol-phosphate mannose synthase depletion in zebrafish Bogenhagen, D.F., et al., 2018. Kinetics and mechanism of mammalian mitochondrial leads to dystrophicmuscle with hypoglycosylated α-dystroglycan. Biochem. Biophys. ribosome assembly. Cell Rep. 22, 1935–1944. Res. Commun. 477, 137–143. Brodersen, D.E., Nissen, P., 2005. The social life of ribosomal proteins. FEBS J. 272, Menezes, M.J., et al., 2015. Mutation in mitochondrial ribosomal protein S7 (MRPS7) 2098–2108. causes congenital sensorineural deafness, progressive hepatic and renal failure and Brown, A., et al., 2014. Structure of the large ribosomal subunit from human mi- lactic acidemia. Hum. Mol. Genet. 24, 2297–2307. tochondria. Science 346, 718–722. Miller, C., et al., 2004. Defective mitochondrial translation caused by a ribosomal protein Bugiani, M., et al., 2004. Clinical and molecular findings in children with complex I (MRPS16) mutation. Ann. Neurol. 56, 734–738. deficiency. Biochim. Biophys. Acta 1659, 136–147. O'Brien, T.W., et al., 2000. Mammalian mitochondrial ribosomal proteins (4). Amino acid Bugiardini, E., et al., 2019. MRPS25 mutations impair mitochondrial translation and sequencing, characterization, and identification of corresponding gene sequences. J. cause encephalomyopathy. Hum. Mol. Genet. https://doi.org/10.1093/hmg/ddz093. Biol. Chem. 275, 18153–18159. Calvo, S.E., et al., 2012. Molecular diagnosis of infantile mitochondrial disease with O'Brien, T.W., et al., 2005. Nuclear MRP genes and mitochondrial disease. Gene 354, targeted next-generation sequencing. Sci. Transl. Med. 4, 118ra10. 147–151. Carroll, C.J., et al., 2013. Whole- sequencing identifies a mutation in the mi- Ott, M., Herrmann, J.M., 2010. Co-translational membrane of mitochondrially tochondrial ribosome protein MRPL44 to underlie mitochondrial infantile cardio- encoded proteins. Biochim. Biophys. Acta 1803, 767–775. myopathy. J. Med. Genet. 50, 151–159. Rizza, T., et al., 2009. Assaying ATP synthesis in cultured cells: a valuable tool for the Chomyn, A., 1996. In vivo labeling and analysis of human mitochondrial translation diagnosis of patients with mitochondrial disorders. Biochem. Biophys. Res. Commun. product. Methods Enzymol. 264, 197–211. 383, 58–62. De Silva, D., et al., 2015. Mitochondrial ribosome assembly in health and disease. Cell Rorbach, J., et al., 2016. Human mitochondrial ribosomes can switch their structural RNA Cycle 14, 2226–2250. composition. Proc. Natl. Acad. Sci. U. S. A. 113, 12198–12201. DeLano, W.L., 2002. The PyMOL Molecular Graphics System. Available. http://www. Saada, A., et al., 2007. Antenatal mitochondrial disease caused by mitochondrial ribo- pymol.org (Accessed on 2018 December 3). somal protein (MRPS22) mutation. J. Med. Genet. 44, 784–786. Desmond, E., et al., 2011. On the last common ancestor and early evolution of : Sali, A., Blundell, T.L., 1993. Comparative protein modelling by satisfaction of spatial reconstructing the history of mitochondrial ribosomes. Res. Microbiol. 162, 53–70. restraints. J. Mol. Biol. 234, 779–815. Distelmaier, F., et al., 2015. MRPL44 mutations cause a slowly progressive multisystem Serre, V., et al., 2013. Mutations in mitochondrial ribosomal protein MRPL12 leads to disease with childhood-onset hypertrophic cardiomyopathy. Neurogenetics 16, growth retardation, neurological deterioration and mitochondrial translation defi- 319–323. ciency. Biochim. Biophys. Acta 1832, 1304–1312. Fernández-Silva, P., et al., 2007. In vivo and in organello analyses of mitochondrial Smith, L.L., et al., 2013. Analysis of skeletal muscle defects in larval zebrafish by bi- translation. Methods Cell Biol. 80, 571–588. refringence and touch-evoke escape response assays. J. Vis. Exp. 82, e50925. Fraher, D., et al., 2016. Zebrafish embryonic lipidomic analysis reveals that the yolk cell is Smits, P., et al., 2007. Reconstructing the evolution of the mitochondrial ribosomal metabolically active in processing lipid. Cell Rep. 14, 1317–1329. proteome. Nucleic Acids Res. 35, 4686–4703. Galmiche, L., et al., 2011. Exome sequencing identifies MRPL3 mutation in mitochondrial Smits, P., et al., 2011. Mutation in mitochondrial ribosomal protein MRPS22 leads to cardiomyopathy. Hum. Mutat. 32, 1225–1231. Cornelia de Lange-like phenotype, brain abnormalities and hypertrophic cardio- Gardeitchik, T., et al., 2018. Bi-allelic mutations in the mitochondrial ribosomal protein myopathy. Eur. J. Hum. Genet. 19, 394–399. MRPS2 cause sensorineural hearing loss, hypoglycemia, and multiple OXPHOS Spirina, O., et al., 2000. Heart-specific splice-variant of a human mitochondrial ribosomal complex deficiencies. Am. J. Hum. Genet. 102, 685–695. protein (mRNA processing; tissue specific splicing). Gene 261, 229–234. Gibert, Y., et al., 2013. Metabolic profile analysis of zebrafish embryos. J. Vis. Exp. 14, Thisse, C., Thisse, B., 2008. High-resolution in situ hybridization to whole-mount zeb- e4300. rafish embryos. Nat. Protoc. 3, 59–69. Gopisetty, G., Thangarajan, R., 2016. Mammalian mitochondrial ribosomal small subunit Tolomeo, D., et al., 2019. Clinical and neuroimaging features of the m.10197G > A (MRPS) genes: a putative role in human disease. Gene 589, 27–35. mtDNA mutation: new case reports and expansion of the phenotype variability. J. Greber, B.J., Ban, N., 2016. Structure and function of the mitochondrial ribosome. Annu. Neurol. Sci. 399, 69–75. Rev. Biochem. 85, 103–132. Torraco, A., et al., 2018. ISCA1 mutation in a patient with infantile-onset leukodystrophy Greber, B.J., et al., 2014a. Architecture of the large subunit of the mammalian mi- causes defects in mitochondrial [4Fe-4S] proteins. Hum. Mol. Genet. 27, 3650. tochondrial ribosome. Nature 505, 515–519. Vacaru, A.M., et al., 2014. Molecularly defined unfolded protein response subclasses have Greber, B.J., et al., 2014b. The complete structure of the large subunit of the mammalian distinct correlations with fatty liver disease in zebrafish. Dis. Model. Mech. 7, mitochondrial ribosome. Nature 515, 283–286. 823–835. Greber, B.J., et al., 2015. Ribosome. The complete structure of the 55S mammalian mi- Vangone, A., et al., 2011. COCOMAPS: a web application to analyze and visualize con- tochondrial ribosome. Science 348, 303–308. tacts at the interface of biomolecular complexes. Bioinformatics 27, 2915–2916. Howarth, D.L., et al., 2013. Defining hepatic dysfunction parameters in two models of Westerfield, M., 2000. The Zebrafish Book. A Guide for the Laboratory Use of Zebrafish fatty liver disease in zebrafish larvae. Zebrafish 10, 199–210. (Danio rerio), 4th edition. University of Oregon Press, Eugene.

11